Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A document search apparatus comprising: a processor that is configured to: acquire from a user an input sentence; perform a morphological analysis of the input sentence to extract words; extract, from the extracted words, a word for use in a search as a query word; estimate a search intention of the user based on the input sentence; acquire, from a search expression list in which a plurality of search expressions are recorded, a search expression corresponding to the estimated search intention; acquire, from an additional query word list in which a plurality of additional query words are recorded, an additional query word corresponding to a set of the acquired search expression and the extracted query word; generate a first query by applying the extracted query word and the acquired additional query word to the acquired search expression; search for a document that matches the first query; and output a result of the search to the user by at least one of a screen display and synthetic speech.
2. The document search apparatus according to claim 1 , wherein if the acquired search expression includes a function to acquire an additional query word, the processor acquires, from the additional query word list, an additional query word corresponding to a set of the function and the extracted query word, and replaces the function in the acquired search expression with the acquired additional query word.
3. The document search apparatus according to claim 1 , wherein the processor is further configured, if there are a plurality of candidates of search expressions corresponding to the estimated search intention, to generate a response sentence to be output to the user to narrow down the candidates.
4. The document search apparatus according to claim 3 , wherein the processor acquires another input sentence input by the user in reply to the response sentence, extracts another query word based on the other input sentence, and generates the first query based on the extracted query word and the extracted other query word.
5. The document search apparatus according to claim 1 , wherein if the user inputs another input sentence, the processor generates a second query based on the other input sentence, generates a third query by connecting the first query and the second query, and searches for a document that matches the third query.
A document search apparatus is designed to improve the accuracy and relevance of search results by dynamically refining queries based on user input. The apparatus includes a processor that generates an initial query from a user's first input sentence. If the user provides additional input, the processor creates a second query from this new input and combines it with the first query to form a third, more comprehensive query. The apparatus then searches for documents that match this refined query, enhancing the search results by incorporating multiple user inputs. This iterative query refinement process allows the system to better understand the user's intent and retrieve more relevant documents. The apparatus may also include a display for presenting the search results to the user, ensuring a seamless and interactive search experience. The system is particularly useful in applications where users need to refine their search criteria incrementally, such as legal research, academic studies, or complex data retrieval tasks. By dynamically adjusting the search query based on user feedback, the apparatus improves search efficiency and accuracy, reducing the need for multiple separate searches.
6. The document search apparatus according to claim 1 , wherein the processor holds a search intention list in which a plurality of search intentions are recorded in association with a keyword or a key phrase, and acquires a search intention corresponding to the input sentence from the search intention list.
7. The document search apparatus according to claim 1 , wherein if the acquired search expression includes a function to acquire an additional query word, the processor acquires, from the additional query word list, an additional query word corresponding to a set of the function and the extracted query word, and generates the first query by applying the extracted query word and the acquired additional query word to the acquired search expression, and if the acquired search expression does not include the function, the processor generates the first query by applying the extracted query word to the acquired search expression.
8. A document search method for a document search apparatus that includes a processor, the document search method comprising, performing by the processor, operations of: acquiring from a user an input sentence; performing a morphological analysis of the input sentence to extract words; extracting, from the extracted words, a word for use in a search as a query word; estimating a search intention of the user based on the input sentence; acquiring, from a search expression list in which a plurality of search expressions are recorded, a search expression corresponding to the estimated search intention; acquiring, from an additional query word list in which a plurality of additional query words are recorded, an additional query word corresponding to a set of the acquired search expression and the extracted query word; generating a first query by applying the extracted query word and the acquired additional query word to the acquired search expression; searching for a document that matches the first query; and outputting a result of the search to the user by at least one of a screen display and synthetic speech.
9. The document search method according to claim 8 , wherein if the acquired search expression includes a function to acquire an additional query word, the acquiring the additional query word comprises acquiring, from the additional query word list, an additional query word corresponding to a set of the function and the extracted query word, and the method further comprises replacing the function in the acquired search expression with the acquired additional query word.
10. The document search method according to claim 8 , further comprising generating, if there are a plurality of candidates of search expressions corresponding to the estimated search intention, a response sentence to be output to the user to narrow down the candidates.
11. The document search method according to claim 10 , further comprising: acquiring another input sentence input by the user in reply to the response sentence; and extracting another query word from the other input sentence; and wherein the generating the first query comprises generating the first query based on the extracted query word and the extracted other query word.
This invention relates to document search systems that improve search accuracy by dynamically refining queries based on user interactions. The problem addressed is the inefficiency of traditional search methods that rely solely on initial user input, often failing to capture the user's true intent or context. The solution involves a multi-step process where an initial query is generated from a user's input sentence, and then refined through subsequent user interactions. The method begins by receiving an input sentence from a user and extracting query words from it to form a first query. This query is used to retrieve relevant documents from a database. The system then generates a response sentence based on the retrieved documents and presents it to the user. If the user provides additional input in response to the initial response, the system extracts query words from this new input and incorporates them into the first query. This iterative refinement process allows the search system to better align with the user's intent, improving the relevance of subsequent search results. The method may also involve ranking the retrieved documents based on their relevance to the refined query, further enhancing search accuracy. This approach is particularly useful in applications where user intent is complex or evolving, such as legal research, medical diagnostics, or customer support systems.
12. The document search method according to claim 8 , further comprising: acquiring from the user another input sentence; generating a second query based on the other input sentence; and generating a third query by connecting the first query and the second query, and wherein the searching comprises searching for a document that matches the third query.
13. The document search method according to claim 8 , wherein the estimating the search intention of the user comprises acquiring, from a search intention list in which a plurality of search intentions are recorded in association with a keyword or a key phrase, a search intention corresponding to the input sentence.
14. The document search method according to claim 8 , wherein if the acquired search expression includes a function to acquire an additional query word, the acquiring the additional query word comprises acquiring, from the additional query word list, an additional query word corresponding to a set of the function and the extracted query word, and the generating the first query comprises generating the first query by applying the extracted query word and the acquired additional query word to the acquired search expression, and if the acquired search expression does not include the function, the generating the first query word comprises generating the first query by applying the extracted query word to the acquired search expression.
15. A non-transitory computer readable medium including computer executable instructions, wherein the instructions, when executed by a processor, cause the processor to perform a method comprising: acquiring from a user an input sentence; performing a morphological analysis of the input sentence to extract words; extracting, from the extracted words, a word for use in a search as a query word; estimating a search intention of the user based on the input sentence; acquiring, from a search expression list in which a plurality of search expressions are recorded, a search expression corresponding to the estimated search intention; acquiring, from an additional query word list in which a plurality of additional query words are recorded, an additional query word corresponding to a set of the acquired search expression and the extracted query word; generating a first query by applying the extracted query word and the acquired additional query word to the acquired search expression; searching for a document that matches the first query; and outputting a result of the search to the user by at least one of a screen display and synthetic speech.
Unknown
February 23, 2021
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.